| 摘要: |
| 本文研究了ARMA-BP神经网络组合模型预测的问题.利用最大最小贴近度评价方法,获得了ARMA-BP神经网络组合模型对应的贴近度差值大于单一模型的结果,推广了组合模型比单一模型预测精度更优的结果. |
| 关键词: ARMA模型 BP神经网络 组合模型 最大最小贴近度 |
| DOI： |
| 分类号:O213 |
| 基金项目:湖北省科技支撑计划软科学研究类项目(2014BDH118). |
|
| FISCAL REVENCE PREDICTION ABOUT THE ARMA-BPNEURAL NETWORK COMBINATION MODEL |
|
FANG Bo, HE Lang
|
|
College of Science, Wuhan University of Technology, Wuhan 430070, China
|
| Abstract: |
| This paper studies the prediction of ARMA-BP neural network combination model. Using the evaluation method of maximum-minimum degree, we get the result that the degree of difierence of ARMA-BP neural network combination model is greater than the single model, which generalizes the conclusion that the combination model is superior to the single model on prediction precision. |
| Key words: ARMA model BP neural network combination model maximum-minimum degree |
